Alternative Heat Sink Ideas!

Got a couple of TWOJ Stars to play with, here are a couple of ideas I have for heat sinks - either temporary or permanent.

1) Silver Half Dollar! Comparitively cheap (I already have it and it's not worth much) and silver is the best heat conductor there is.

2) Copper pipe end cap. The cap for 1" pipe is just about the right size for the star, I'm thinking about maybe putting the star in the bottom of the cap, in which case the optics / reflector would be nicely protected by the sides of the cap.

3) Spot Welding Tip. Imagine a 1"x1/2" copper cylider. 1/2" diameter, with the top 1/4" tapering to 1/4" diameter. Of course, if you had to buy these, I don't think they'd be very cheap. I happen to have one sitting around...

If I grind down the tip to perhaps 3/8", an emitter should fit very nicely. It'll sit on at least 1/2" solid copper before it becomes a hollow copper tube (for cooling water on spot welders)

I saw some brass plugs for plumbing, but they were quite a bit more than the copper pipe caps. Also, brass isn't nearly as good a conductor as copper.

Any other ideas for cheap, easy to find alternative heat sinks?

I like the 1/2 dollar idea, am old quarter(silver, not exactially sure what year they stopped using silver) could also work as an inbetween peice.
nice idea '262'
Jeff

for emitters, I've begun using pre 1982 pennies (more Cu in them as someone else pointed out).

last year for silver quarters and dimes was 1964

when observing a heat transfer chart thing
there is a big jump from
steel to aluminum
aluminum to copper
but there isnt much MORE jump to silver, and other high priced items.

although silver would be just plain Neeto

and another one of my ideas on heatsinks
copper MOVES heat all about the place
at some point in all this movement stuff
you need to DISSIPATE the heat to the air
and the worser heat transfer metals do exactally that
they dont MOVE it as well, as they toss it off
so
it would be preferable to use copper to move or spread it
but something else to get RID of it into the air.
like the head of the light for example.

that is why a great heat sync uses a copper spreader
but aluminum fins (assuming they made good contact)

Sorry, but I really don't remember enough of my ME109 (heat transfer) to understand the diff between heat movement and dissipation. Is there a link I can look at? I saw this somewhere else too...

In any heat transfer process, you have set of consecutive pathways that heat must travel through from the heat source (in this case, the LED die) to its final exhaust point (in this case, the ambient air). For a well-heatsinked LED flashlight, the largest thermal resistance will be in getting the heat from the body of the flashlight to the air. Improvements to the LED-heatsink contact, or to the heatsink material itself, will quickly encounter diminishing returns unless either of these is poorly designed to begin with.

These thermal resistances work in a similar manner to electrical resistors in series. You are unlikely to get a useful reduction of the total resistance by making the smallest resistor even smaller.

i dont have a link
i just dream this stuff up from logic
and from experimenting with the different stuff.

if something doesnt transfer heat well, where does the heat go? unless its a thermal insulator.

take a copper bar in the AIR, heat one side , the other side begins to get warmer.
take a aluminum bar in the air heat one side, and in the other side doesnt get as warm.
where did the heat go?
into the air , i assume.

I looked up dissipation, and it's actually about the same between CU and AL. Discussing this with some other engineers here, so far we're very skeptical. Still, I saw a computer site where someone was saying the same thing about heat sinking CPU's.

Vid - as far as the difference in heat, that can be explained by the transfer rate of heat into the bar - at an extreme, if you used a wood heatsink, the other side would be a much lower temperature. What is key is the temp at the heat source itself. I'm still trying to figure this one out...

Dissipation as I understand it has to do with emissivity, which, for lack of better terms, is the thermal bond between your solid object and the "luminous ether;" how well thermal wavelengths jump from the solid object.

Roughly spreaking, the more reflective the object, the lower (I think, or maybe higher -- always get that backward) the emissivity. Translation: shiny things neither get hot in the sun nor cool off well. They tend not to release those "therms."

This is why space suits are white and wood stoves are black (yes, the steam radiator in your old dorm room would have worked better black.)

You can boost copper's emissive characteristics (heat shedding) by blackening it with something like gun blueing or some other oxidized coating. The coatings themselves aren't usually very thermally conductive, but they are thin enough that emissivity becomes more important.

Scott

Liquid Cooled Luxeon?

Two definitions that may be useful in this thread:

Thermal Conductivity is the ease with which a material moves heat about inside itself.

Thermal Capacity is the ability of a material to soak up and contain heat.

One of these days I'd like to see what happens if you run a luxeon submerged in liquid, say mineral oil. This would allow you to get the front and sides of the emitter some cooling, and hopefully get some convection going.

Re: Liquid Cooled Luxeon?

oops, i try to not use technical terms
more like standard english definitions instead.
that way its simplified.

my use of the word was more along these lines
dissipate (dîs´e-pât´) verb
dissipated, dissipating, dissipates verb, transitive
1. To drive away; disperse.
2. To attenuate to or almost to the point of disappearing: The wind finally dissipated the smoke. See synonyms at scatter.
3. a. To spend or expend intemperately or wastefully; squander. b. To use up, especially recklessly; exhaust: dissipated their energy. See synonyms at waste.
4. To cause to lose (energy, such as heat) irreversibly.


dont ask me to get technical its not my job.

if the only concideration is transfer of heat from the one hot spot, then copper works great, till the whole copper is blazing hot, cause it didnt go anywere to leave.

[ QUOTE ]
beezaur said:
Dissipation as I understand it has to do with emissivity, which, for lack of better terms, is the thermal bond between your solid object and the "luminous ether;" how well thermal wavelengths jump from the solid object.


[/ QUOTE ]
Basically correct except that emissivity plays a very small part in getting rid of heat when there is conduction and convection. In a vacuum it is the only way heat is emitted, but the types of heat sinks being discussed here mostly conduct their heat to the air. Also, emission goes up in proportion to absolute temperature to the fourth power, so it does play a role when something is very hot, like a light bulb filament, but not with things at or near room temperature.

[ QUOTE ]

Roughly spreaking, the more reflective the object, the lower (I think, or maybe higher -- always get that backward) the emissivity. Translation: shiny things neither get hot in the sun nor cool off well. They tend not to release those "therms."


[/ QUOTE ]
Again if you're talking about getting rid of the heat, then this is only true in a vacuum. Absorbing the heat is another story, but that has more to do with the albedo of the object in the type of light it is in. If an object absorbs most wavelengths (i.e. black), then it will tend to get hotter. If it reflects most wavelengths (white or silver) it will tend to not get as hot. BTW, those old-fashioned radiators work mostly by convection, so the color is immaterial. They're usually silver because the aluminum in silver paint acts as rust-proofing.

[ QUOTE ]

You can boost copper's emissive characteristics (heat shedding) by blackening it with something like gun blueing or some other oxidized coating.

[/ QUOTE ]
Theoretically, yes. Practically, it'll make less than 10% difference with a passive heat sink and negligible difference with a fan-forced one. Also, a good part of the improvement in the passive mode has nothing to do with blackening but rather is due to the fact that the prepartion for anodizing and/or blueing roughens up the surface of the heat sink. This improves performance by increasing surface area.

Two things are important with heat sinks as far as getting rid of the heat. You need surface area first and air flow second. A passive heat sink will set up convective air flow as it gets warm. A fan-forced one depends upon a fan, but has an order of magnitude better performance for the same size. Using a fluid with better conductivity than air improves performance yet again. I made liquid heat sinks out of copper which can deal with 200W from thermoelectric modules while rising only about 3°C above the temperature of the water. And they're only 1.625" x 4" x 0.5" to boot. Such performance would be almost impossible with air cooling.

Re: Liquid Cooled Luxeon?

The thermal path in an LS starts from the die to the slug. Ain't nothin we can do to improve that. /ubbthreads/images/graemlins/frown.gif The next path is the most important, due to the small surface area involved- the slug-to-heatsink. Compared to metals, thermal pastes/epoxies are terrible conductors of heat! Their coefficients of thermal conductivity (CTC) are a small fraction of those for metals.

Conductive thermal transfer between the slug and heatsink is a function of 4 things:
1. The surface area of the slug (Unless your sink's area is smaller. /ubbthreads/images/graemlins/icon15.gif )
2. The temperature difference between the slug and heatsink at the area of contact. (No difference, no flow.)
3. The thermal conductivity of the paste. (Numerically terrible!)
4. The thickness of the paste layer. (Goes in the denominator of the transfer equation- thicker = worse.)

I remember Don/McGizmo farting around with diamond (Highest thermal conductivity) dust in attempts to make better thermal goo! Phonon flow! /ubbthreads/images/graemlins/icon15.gif (Search Home-made and modified, where this subject has been covered in depth several times.)

Lacking any means to increase the slug's surface area, the best hope is to reduce the thickness of the interface layer. The slug and mating area on the sink should be as flat and smooth as possible!!! This permits the thinnest layer of interface. I lap both slugs and sink contact when possible. Most computer overclockers lap their sinks also.

BTW, the primary means of extracting heat from small hot flashlights is by conduction to the user's hand. I found I could hold this one continuously, but heaven forbid setting it down for a while and returning for it. /ubbthreads/images/graemlins/grin.gif As my vacationing friend who (fortunately) can't reply would say, /ubbthreads/images/graemlins/nana.gif










Larry

Re: Liquid Cooled Luxeon?

well he initially said STAR :-( just to confuse us.
i soldered an emitter on, are you saying that the lead would be a better conductor than some wizz bang arctic silver epoxy junk (very thin)?

i dont recommend soldering them . . . of course.

"Again if you're talking about getting rid of the heat, then this is only true in a vacuum. Absorbing the heat is another story, but that has more to do with the albedo . . ."

Albedo, as in "ratio of reflected to incident light?"

Why is it true only in a vacuum? Air is transparent to "thermal" wavelengths.

"prepartion for anodizing and/or blueing roughens up the surface of the heat sink. This improves performance by increasing surface area. . . ."

I am not sure what you mean by this. The surface area will be greatly ncreased by roughening, but only on tiny scales. I don't see how that would be significant to conduction or convection. Even radiation would seem to be little effected by that type of surface area increase.

I am very surprised at 10% difference between emissivity near 1 vs. near 0. I have noticed changes in cooling half-times under ambient conditions that I would place closer to 50% or higher. This is not working with electronics (small structures), but the principles are the same.

It depends on what your device is "looking at" too. If you are experimenting with something inside a metal box, or something in proximity to other radiating bodies, the situation favors conduction/convection.

Scott

Re: Liquid Cooled Luxeon?

Hi Vid, I missed the star part /ubbthreads/images/graemlins/frown.gif . Yes, soldering a slug should be very superior to Arctic Silver paste! (Google "coefficients of thermal conductivity") It's been done before and a lot of us went /ubbthreads/images/graemlins/faint.gif . I haven't had the huevos to try it! One concern is the LS is going to "float" on the molten solder, so minimizing the amount/starting bead would be benificial. As I recall, the Arctic folks are very honest and list the thermal conductivity of their products. (I Like and use their products!)

Larry

[ QUOTE ]
beezaur said:
Albedo, as in "ratio of reflected to incident light?"


[/ QUOTE ]
Yes, exactly.

[ QUOTE ]

Why is it true only in a vacuum? Air is transparent to "thermal" wavelengths.


[/ QUOTE ]
It works in air too (remember the heat you feel from incandescents?), but the amount of heat lost due to convection and conduction is usually several orders of magnitude greater so emission is negligible.

[ QUOTE ]

I am not sure what you mean by this. The surface area will be greatly ncreased by roughening, but only on tiny scales. I don't see how that would be significant to conduction or convection. Even radiation would seem to be little effected by that type of surface area increase.


[/ QUOTE ]
On a microscopic scale the surface area is increased by 10% or so. The surface texture of the object might not look a whole lot different than it did before, but it does have a measureable effect on heat transfer.

[ QUOTE ]

I am very surprised at 10% difference between emissivity near 1 vs. near 0. I have noticed changes in cooling half-times under ambient conditions that I would place closer to 50% or higher. This is not working with electronics (small structures), but the principles are the same.


[/ QUOTE ]
Changes of 50% mean that the device being cooled only rises half as much above ambient as before. Do you have any numbers you could share? Rise above ambient for a given power dissipation is the only measure of a heat sink's efficiency. The usual units for this are °C/W (lower is better). Blackening cuts this number by up to 10%.

Re: Liquid Cooled Luxeon?

beezaur, yes, conduction rules! Conduction occurs at the micro/submicro level for each transfer interface. This is valid from the die/slug/star board/heatsink/flashlight body/user's (liquid-cooled) hand/atmosphere through a lot of skin area/latent heat of sweat evaporation/global warming. /ubbthreads/images/graemlins/grin.gif

Larry

I have a very large heatsink on my Athlon 64 3000+ (Winchester core) A Thermalright XP90 beast, it has a copper core, aluminum fins and four heat pipes. A 92mm can fan slowly spins keeping my CPU at 41C surfing and 48C benchmarking. (The Panaflo L series fan is running at 1307 RPM at 5 volts and is silent at 30cm away) /ubbthreads/images/graemlins/buttrock.gif

Get some small heat pipes and mount it in copper. Have the heat pipes run the length of the flashlight in the aluminum barrel.

Don't ask me were to get tiny heat pipe diameter stuff, I just come up with the ideas! /ubbthreads/images/graemlins/rolleye11.gif Actually, the heat pipe thing would work very well with flashlights. A 5 Luxeon III beast would be a great hand warmer in the winter.